Boost converter how?

[Pommie]

I've ordered some IRL540s to play with and I had assumed that I could monitor the battery current and alter the PWM to maximise it. Wouldn't this find the max power point? It would actually be easier to regulate the input voltage but I assume that this would limit you to a certain type of panel or is this the same for all panels?

Mike.

 

[john warfin]

It's still not worth making a buck switcher for 12v to 12v system.

Hello. I'm an EE experienced in switcher design but new to the solar power arena. Having built a couple MPPT have come to the same conclusion regarding 12v/12v so I hope to benefit from advice of those here more experienced in the area than myself.

I understand benefits of 12v solar on 24v batteries using boost. However for small (100-200w) systems consumer 12v battery and inverter seems most economical.

Currently using four 22v peak 30w amorphous panels in parallel with a buck converter. Would there be advantages to reconfiguring for 44v into the 12v battery?

Will one or the other make more effective use of morning and late afternoon sun? Or will going back to PWM make more sense. Any opinions?

 

[smanches]

It really has to do with the panel output voltage during these low light times. If it is close to the battery voltage already, then MPPT isn't going to do a whole lot.

In general, running your panels at higher voltages would be the most efficient. You'll have less I2R losses getting to the controller, and more voltage headroom for the controller to play with during the lower light scenarios. This all assumes you're using MPPT though. Standard PWM controllers should be run as close to the battery voltage as possible to gain the most power.

 

[Mr RB]

I've ordered some IRL540s to play with and I had assumed that I could monitor the battery current and alter the PWM to maximise it. Wouldn't this find the max power point? It would actually be easier to regulate the input voltage but I assume that this would limit you to a certain type of panel or is this the same for all panels?
...

Technically a MPPT should adjust to get max power into the battery. But it's not necessary, since the panel peak power output is usually around the same voltage even for different insolation levels, so even at low light levels the panels will still make max power near 17v. So you can regulate for input voltage which is much easier and gets you very close. To do it the other system is more complex, and you need to text the max power output by "detuning" the system away from maximum, to find where the max is, then tune it back to maximum. That adds some losses to the complexity.

As for different panels, you said you already had the panel. So tune it for that. If you are making a device for general use or to sell then stay away from boost, just use 24v panels to charge 24v batteries.

 

[john warfin]

So you can regulate for input voltage which is much easier and gets you very close./QUOTE]

If done properly doesn't that require sensing light level and complex algorithms? I've been maximizing battery current which involves neither. Just a small extra wire from micros ADC to battery.

 

[Mr RB]

So you can regulate for input voltage which is much easier and gets you very close.

If done properly doesn't that require sensing light level and complex algorithms?

I haven't done much testing with the cheaper amorphous panels but they are rarely used for MPPT use anyway as their max power is delivered closer to 15v than 17v.

But with the crystalline type that produce their max at 17 or so volts this Vin is fairly constant. The shape of the V/I curve (with max power produced at about 17v) is tied to the physical properties of the panel which remain pretty constant. If you reduce the light level the current output drops but the V/I curve stays pretty much the same. So it still makes around 22v open circuit in full sun and 22v open circuit in cloud. Likewise the max power out in full sun is about 17v and the max power out when cloudy is still about 17v.

When I was working on the commercial MPPT design I had two 12v 80W panels setup for a few days, and I run them into a high power adjustable load and test meters. The max power out was always around the same panel voltage in cloud or full sun etc, and I looked into the option of making a MPPT that just regulated input voltage which is a valid system provided you have an adjuster so set the input voltage when it is first set up and tested. (The peak power point may be different in VERY low light levels but the panel only produces an insignificant % of its total power under those light conditions anyway.)

Testing current into the battery is an excellent system because it auto adjusts for things like the solar panel cable losses and SMPS factors etc. But to "track" the max you need to detune away from max, in both directions, regularly to keep checking that it was ON the max. There is very little difference in total long term power output between a properly adjusted Vin regulated SMPS and a tracking SMPS, so for any small simple system (like Pommie's) I would just go with regulated Vin.

 

[john warfin]

So it still makes around 22v open circuit in full sun and 22v open circuit in cloud. Likewise the max power out in full sun is about 17v and the max power out when cloudy is still about 17v.

Unloaded voltage makes perfect sense to me however the loaded statement seems somewhat counter-intuitive. Why are these guys spending so much time and money on light sensing sensing systems to stay on the cliff edge? It should be easy to verify with a couple simple experiments.

Apparently I've got a lot to learn.

 

[MrAl]

Hi there, I have boost converter software that helps during the compensation network design phase that was developed as a result of finding errors upon errors on articles found on the web about compensating a boost converter. If you are interested i could send you a copy. It takes into consideration just about everything like inductor esr, capacitor esr, switch resistance, input source resistance, etc., and unlike other types of software this works in the time domain so you can see the results of a component selection in the same way you would if you built a converter and used that component. It also runs very fast (under 1 second per run) because it is made from a set of pre solved equations for a boost converter. The only requirement is that it runs under Windows.

 

[AllVol]

Hi,
Just done a test download, all ok.

Using the same zipper for other members posts, no reported problems.???

I get "invalid or corrupted file" box when I try to download.

 

[Mr RB]

Unloaded voltage makes perfect sense to me however the loaded statement seems somewhat counter-intuitive. Why are these guys spending so much time and money on light sensing sensing systems to stay on the cliff edge? It should be easy to verify with a couple simple experiments.
...

Light sensing is completely irrelevant for MPPT! Light sensing is used for tracking the angular position of the sun, to mechanically move the panels (solar tracking) to point them in the direction where they get the most light.

Max power point tracking cares nothing about light it simply adjusts the Vin of the SMPS to provide the maximum power output from the SMPS. It's an adjustable gearbox, that self-adjusts the gear ratio to give the most power output.

As for the simple experiments, yes I did those, like I said I run 2 80W panels for some days into adjustable load and tested power point under varying conditions.

 

[john warfin]

Light sensing is completely irrelevant for MPPT! Light sensing is used for tracking the angular position of the sun, to mechanically move the panels (solar tracking) to point them in the direction where they get the most light.

According to a couple sources I downloaded light sensing is the key to successful MPPT. Either by momentarily switching the main panel to a precision resistor and sensing voltage. Or eliminating the high current switch and using a smaller "reference" PV cell with fixed load. In either case a fixed reference of some kind was needed when measuring voltage to see where it is on the curve. Then high climb (hill climb) algorithms went into effect.

I tried both with my early MPPT designs but finally settled on a battery current method simply because it did not require a light sensor. You are saying it's possible to use the MPPT controller itself as the reference load?

PS I've also designed and built two LED array trackers loosely based in the Redrok unit so do know the difference compared to power point tracker.

 

[Mr RB]

Light sensing is irrelevant. The one and only factor to be maximised is current into the battery. (Power into the battery is irrelevant too, if you think about it). No fixed reference is needed!

So the best system is one that monitors current into the battery and adjusts SMPS PWM to get the most current into the battery.

What I was saying is that for a small SIMPLE system you can rely on the solar panel properties remaining fairly constant through most insolation levels, so you can "cheat" by just regulating the SMPS Vin to keep the panel voltage at the voltage where that panel makes peak power. Of course if the system is larger it's worth the extra parts and complexity of making it closed loop and monitoring current into the battery.

 

[john warfin]

The one and only factor to be maximised is current into the battery.

No fixed reference is needed!

So the best system is one that monitors current into the battery and adjusts SMPS PWM to get the most current into the battery.

That is exactly the conclusion I came to.

What I was saying is that for a small SIMPLE system you can rely on the solar panel properties remaining fairly constant through most insolation levels, so you can "cheat" by just regulating the SMPS Vin to keep the panel voltage at the voltage where that panel makes peak power. Of course if the system is larger it's worth the extra parts and complexity of making it closed loop and monitoring current into the battery.

Yes. Adjusting duty cycle to maximize battery current does however result in significant changes in panel voltage under different lighting conditions with my amorphous cells. I consider this closed loop system worthwhile for even my small setup as apposed to just hooking arrays up to a battery. About 20% or so advantage so far.

I am now trying to determine if putting panels in series is more appropriate for my 12v battery/inverter to take more advantage of low light levels. I'm more into experimentation than theory so maybe I'll just try it.